The definition of preevent soil moisture profile and time compression analysis are critical components in water balance models that are based on realistic infiltration/exfiltration relations and include profile redistribution of vadose zone moisture. In this paper, detailed analysis of these two fundamental components of water balance modeling is presented. Numerical integration of the governing equations for liquid moisture flow in the unsaturated zone is used in simulations designed to illustrate the role of temporal variability in the system. The simulations consist of forcing the surface of a one-dimensional soil column, bounded at its base by a fixed water table, with the output of a stochastic event-based model of precipitation and potential evaporation. The simulations are run until (and beyond when) an equilibrium condition is reached between the long-term mean values of surface and bottom fluxes. For situations with a deep water table, two distinct zones develop: a near-surface, highly unsteady zone, and a deeper, quasi-steady zone. The equivalent steady moisture profile, i.e., the steady profile corresponding to the mean of the simulated column flow, is found to reasonably approximate the temporal mean, mean prestorm, and mean poststorm moisture profiles, particularly in the deeper zone. In the upper zone the equivalent steady profile forms a biased estimate of the temporal mean.

The bias is shown, through perturbation analysis, to lead to overestimation of the wetness of the mean moisture profile. For shallow water tables the distinction between the two zones collapses. In this case both the equivalent steady profile and the mean profile are close to the hydrostatic profile. The simulations are also used to test the utility of the time compression approximation (TCA) in modeling surface fluxes under temporally variable initial conditions. We demonstrate that the use of the mean prestorm and poststorm event moisture profiles as initial condition and exfiltration flux capacities does not produce major bias in TCA-predicted surface fluxes. In this sense the TCA behaves as a linear operator. Furthermore, we demonstrate that the estimation of these mean initial condition profiles by the equivalent steady state soil moisture profile is an adequate approximation for determining mean land surface response to event-based atmospheric forcing. ¿ American Geophysical Union 1994